Global Situation

Child abuse and neglect are not rare and occur worldwide. WHO has estimated that 40 million children under the age of 15 yr suffer from abuse and neglect; it has been estimated that over 1 million children have been trafficked. Studies from many countries across the globe indicate that more than 80% of children suffer physical punishment in their homes, with over 30% experiencing severe physical punishment. Surveys reported by UNICEF confirm these reports; 1 survey conducted in the Middle East reported that 30% of children had been beaten or tied up by parents, and in a 2nd survey in a Southeast Asian country, 30% of mothers reported having hit their child with an object in the past 6 mo.

United States Situation

In the USA, abuse and neglect mostly occur behind closed doors and often are a well kept secret. Nevertheless, there were over 3 million reports to CPS involving over 6 million children in the USA in 2008—a rate of 49.4 per 1000 children. Of the 772,000 substantiated reports, 71% were for neglect, 2.2% for medical neglect, 16% for physical abuse, 9% for sexual abuse, and 7.3% for psychologic maltreatment. This translates to a rate of 8 per 1000 children identified as neglected, a rate that has been steady since the early 1990s; this is in contrast to declining rates of sexual abuse (down 53%) and physical abuse (down 48%) between 1992 and 2006. Medical personnel make about 12% of all reports.

The highest rate of maltreatment, 16.4 per 1000, was for children 0 to 3 yr old. The rate for black children was almost twice that of white children (20.4 vs 11.0 per 1000). This rate is likely due, in part, to professional bias in identifying, reporting, and investigating low-income and minority families for maltreatment. Sources independent from the official CPS statistics cited above confirm the prevalence of child maltreatment. In a community survey, 3% of parents reported using very severe violence (e.g., hitting with fist, burning, using gun or knife) against their child in the prior year. Considering a natural disinclination to disclose socially undesirable information, such rates are both conservative and alarming.

Etiology

Child maltreatment seldom has a single cause; rather, multiple and interacting biopsychosocial risk factors at 4 levels usually exist. At the individual level, a child’s disability or a parent’s depression or substance abuse predispose a child to maltreatment. At the familial level, intimate partner (or domestic) violence presents risks for children. Influential community factors include stressors such as dangerous neighborhoods or a lack of recreational facilities. Professional inaction may contribute to neglect, such as when the treatment plan is not clearly communicated. Broad societal factors, such as poverty and its associated burdens, also contribute to maltreatment. WHO estimates the rate of homicide of children is approximately twofold higher in low-income compared to high-income countries (2.58 vs 1.21 per 100,000 population), but clearly homicide occurs in high-income countries (Table 37-1). Children in all social classes can be maltreated, and child health care professionals need to guard against biases concerning low-income families.

Table 37-1 CHILD MALTREATMENT DEATHS BY NATION

* Deaths include obvious maltreatment and those of undetermined intent.

From UNICEF: A league table of child maltreatment deaths in rich nations. In Inocenti Report Card No 5, Florence, September 2003, UNICEF Innocenti Research Centre, Figure 1b, p 4.

In contrast, protective factors, such as family supports, or a mother’s concern for her child, may buffer risk factors and protect children from maltreatment. Identifying and building on protective factors can be vital to intervening effectively. One can say to a parent, for example, “I can see how much you love ____. What can we do to keep her out of the hospital?” Child maltreatment results from a complex interplay among risk and protective factors. A single mother who has a colicky baby and who recently lost her job is at risk for maltreatment, but a loving grandmother may be protective. A good understanding of factors that contribute to maltreatment, as well as those that are protective, should guide an appropriate response.

Clinical Manifestations

Child abuse and neglect can manifest in many different ways (Fig. 37-1). With regard to physical abuse, a critical element is the lack of a plausible history other than inflicted trauma. As with any medical condition, the onus is on the clinician to carefully consider the differential diagnosis and not jump to conclusions.

Figure 37-1 The spectrum of child maltreatment. Child maltreatment encompasses acts of commission, abuse, and acts of omission or neglect by a caretaker that adversely affect children. The act can be physical or psychologic. The boundaries between these areas are indistinct and psychologic; physical abuse and neglect overlap and may exist at the same time or various times in a child’s life. Sexual abuse may be considered a specific type of physical abuse that has strong emotional components. Physical abuse and neglect invariably have short- and long-term psychologic consequences. Psychologic consequences may persist long after the physical wounds heal.

Bruises are the most common manifestation of physical abuse. Features suggestive of inflicted bruises include (1) bruising in a preambulatory infant (occurring in just 2% of infants), (2) bruising of padded and less exposed areas (buttocks, cheeks, under the chin, genitalia), (3) patterned bruising or burns conforming to shape of an object or ligatures around the wrists (Figs. 37-2 and 37-3), and (4) multiple bruises, especially if clearly of different ages. Estimating the age of bruises needs to be done cautiously. Red suggests less than a week, yellow suggests more than 1-2 days. It is very difficult to precisely determine the ages of bruises.

Figure 37-2 A variety of instruments may be used to inflict injury on a child. Often the choice of an instrument is a matter of convenience. Marks tend to silhouette or outline the shape of the instrument. The possibility of intentional trauma should prompt a high degree of suspicion when injuries to a child are geometric, paired, mirrored, of various ages or types, or on relatively protected parts of the body. Early recognition of intentional trauma is important to provide therapy and prevent escalation to more serious injury.

Figure 37-3 Marks from heated objects cause burns in a pattern that duplicates that of the object. Familiarity with the common heated objects that are used to traumatize children facilitates recognition of possible intentional injuries. The location of the burn is important in determining its cause. Children tend to explore surfaces with the palmar surface of the hand and rarely touch a heated object repeatedly or for a long time.

Other conditions such as birthmarks and Mongolian spots can be confused with bruises and abuse. These skin markings are not tender and do not rapidly change color or size. An underlying medical explanation for bruises may exist, such as blood dyscrasias or connective tissue disorders (hemophilia, Ehlers-Danlos). The history or examination usually provides clues to these conditions. Henoch-Schönlein purpura, the most common vasculitis in young children, may be confused with abuse. The pattern and location of bruises caused by abuse are usually different from those due to a coagulopathy. Noninflicted bruises are characteristically anterior and over bony prominences, such as shins and forehead. The presence of a medical disorder does not preclude abuse.

Cultural practices can cause bruising. Cao gio, or coining, is a Southeast Asian folkloric therapy. A hard object is vigorously rubbed on the skin, causing petechiae or purpura. Cupping is another approach, popular in the Middle East. A heated glass is applied to the skin, often on the back. As it cools, a vacuum results, leading to perfectly circular bruises. The context here is important, and such circumstances should not be considered abusive.

A careful history of bleeding problems in the patient and first degree relatives is needed. If a bleeding disorder is suspected, a platelet count, prothrombin time, international normalized ratio (the ratio of the prothrombin time to a control sample, raised to the power of the International Sensitivity Index), and partial thromboplastin time should be obtained (Chapter 469). More extensive testing should be considered in consultation with a hematologist.

Bites have a characteristic pattern of 1 or 2 opposing arches with multiple bruises (see Fig. 37-2). They can be inflicted by an adult, another child, an animal, or the patient. Bites by a child (younger than approximately 8 yr with primary teeth) typically have a distance of less than 2.5 cm between the canines—often the most prominent bruises. The appearance of animal bites is variable (Chapter 705); they usually have narrower arches than human bites and are often deep. Self-inflicted bites are on accessible areas, particularly the hands. Adult bites raise concern for abuse. Multiple bites by another child suggest inadequate supervision and neglect.

Burns may be inflicted or due to inadequate supervision. Scalding burns may result from immersion or splash (Fig. 37-4; also see Fig. 37-3). Immersion burns, when a child is forcibly held in hot water, show clear delineation between the burned and healthy skin and uniform depth (see Fig. 37-4). They may have a sock or glove distribution. Splash marks are usually absent, unlike when a child inadvertently encounters hot water. Symmetrical burns are especially suggestive of abuse as are burns of the buttocks and perineum. Although most often accidental, splash burn may also result from abuse. Burns from hot objects such as curling irons, radiators, steam irons, metal grids, hot knives, and cigarettes leave patterns representing the object. A child is likely to try to escape from a hot object; thus burns that are extensive and deep reflect more than fleeting contact and are suggestive of abuse.

Figure 37-4 A 1 yr old child brought to a hospital with a history that she sat on a hot radiator. Suspicious injuries such as this require a full medical and social investigation including a skeletal survey to look for occult skeletal injuries and a child welfare evaluation.

Several conditions mimic abusive burns, such as brushing against a hot radiator, car seat burns, hemangiomas, and folk remedies such as moxibustion. Impetigo may resemble cigarette burns. Cigarette burns are usually 7-10 mm across, whereas impetigo has lesions of varying size. Noninflicted cigarette burns are usually oval and superficial.

Neglect frequently contributes to childhood burns (Chapter 68). Children, home alone, may be burned in house fires. A parent taking drugs may cause a fire and may be unable to protect a child. Exploring children may pull hot liquids left unattended onto themselves. Liquids cool as they flow downward so that the burn is most severe and broad proximally. If the child is wearing a diaper or clothing, the fabric may absorb the hot water and cause burns worse than otherwise expected. Some circumstances are difficult to foresee, and a single burn resulting from a momentary lapse in supervision should not automatically be seen as neglectful parenting.

Concluding whether a burn was inflicted depends on the history, burn pattern, and the child’s capabilities. A delay in seeking health care may result from the burn initially appearing minor, before blistering or becoming infected. This circumstance may represent reasonable behavior and should not be automatically deemed neglectful. A home investigation is often valuable (e.g., testing the water temperature).

Fractures that strongly suggest abuse include: classic metaphyseal lesions, posterior rib fractures, and fractures of the scapula, sternum, and spinous processes, especially in young children (Table 37-2, Fig. 37-5). These fractures all require more force than would be expected from a minor fall or routine handling and activities of a child. Rib and sternal fractures rarely result from cardiopulmonary resuscitation, even when performed by untrained adults. In abused infants, rib, metaphyseal, and skull fractures are most common. Femoral and humeral fractures in nonambulatory infants are also very worrisome for abuse. With increasing mobility and running, toddlers can fall with enough rotational force to cause a spiral, femoral fracture. Multiple fractures in various stages of healing are suggestive of abuse; nevertheless, underlying conditions need to be considered. Clavicular, femoral, supracondylar humeral, and distal extremity fractures in children older than 2 yr are most likely noninflicted unless they are multiple or accompanied by other signs of abuse. Few fractures are pathognomonic of abuse; all must be considered in light of the history.

Figure 37-5 A, Metaphyseal fracture of the distal tibia in a 3 mo old infant admitted to the hospital with severe head injury. There is also periosteal new bone formation of the tibia, perhaps from previous injury. B, Bone scan of same infant. Initial chest x-ray showed a single fracture of the right posterior 4th rib. A radionuclide bone scan performed 2 days later revealed multiple previously unrecognized fractures of the posterior and lateral ribs. C, Follow-up radiographs 2 wk later showed multiple healing rib fractures. This pattern of fracture is highly specific for child abuse. The mechanism of these injuries is usually violent squeezing of the chest.

The differential diagnosis includes conditions that increase susceptibility to fractures, such as osteopenia and osteogenesis imperfecta, metabolic and nutritional disorders (e.g., scurvy, rickets), renal osteodystrophy, osteomyelitis, congenital syphilis, and neoplasia. Features of congenital or metabolic conditions associated with nonabusive fractures include family history of recurrent fractures after minor trauma, abnormally shaped cranium, dentinogenesis imperfecta, blue sclera, craniotabes, ligamentous laxity, bowed legs, hernia, and translucent skin. Subperiosteal new bone formation is a nonspecific finding seen in infectious, traumatic, and metabolic disorders. In young infants, new bone formation may be a normal physiologic finding, usually bilateral, symmetric, and less than 2 mm in depth.

The evaluation of a fracture should include a skeletal survey in children less than 2 yr of age when abuse seems possible. Multiple films with different views are needed; “babygrams” (1 or 2 films of the entire body) should be avoided. If the survey is normal, but concern for an occult injury remains, a radionucleotide bone scan should be performed to detect a possible acute injury (see Fig. 37-5). Follow-up films after 2 wk may also reveal fractures not apparent initially (see Fig 37-5).

In corroborating the history and the injury, the age of a fracture can only be crudely estimated (Table 37-3). Soft-tissue swelling subsides in 2-21 days. Periosteal new bone is visible within 4-21 days. Loss of definition of the fracture line occurs between 10-21 days. Soft callus can be visible after 10 days and hard callus between 14-90 days. These time frames are shorter in infancy and longer in children with poor nutritional status or a chronic underlying disease. Fractures of flat bones such as the skull do not form callus and cannot be aged.

Abusive head trauma (AHT) results in the most significant morbidity and mortality. Abusive injury may be caused by direct impact, asphyxia, or shaking. Subdural hematomas (Fig. 37-6), retinal hemorrhages (especially when extensive and involving multiple layers) (Fig. 37-7), and diffuse axonal injury strongly suggest AHT, especially when they co-occur (Chapter 63). The poor neck muscle tone and relatively large heads of infants make them vulnerable to acceleration-deceleration forces associated with shaking, leading to AHT. Children may lack external signs of injury, even with serious intracranial trauma. Signs and symptoms may be nonspecific, ranging from lethargy, vomiting (without diarrhea), changing neurologic status or seizures, and coma. In all preverbal children, an index of suspicion for AHT should exist when children present with these signs and symptoms.

Figure 37-6 CT scan indicating intracranial bleeding. A, Older blood. B, New blood.

Figure 37-7 Retinal hemorrhages. Lines point to hemorrhages of various sizes.

Acute intracranial trauma is best evaluated via initial and follow-up CT. MRIs are helpful in differentiating extra axial fluid, determining timing of injuries, assessing parenchymal injury, and identifying vascular anomalies. MRIs are best obtained 5-7 days after an acute injury. Glutaric aciduria type 1 can present with intracranial bleeding and should be considered. Other causes of subdural hemorrhage in infants include arteriovenous malformations, coagulopathies, birth trauma, tumor, or infections. When AHT is suspected, injuries elsewhere—skeletal and abdominal—should be ruled out.

Retinal hemorrhages are an important marker of AHT (see Fig. 37-7). Whenever AHT is being considered, a dilated indirect ophthalmologic examination by a pediatric ophthalmologist should be performed. Although retinal hemorrhages can be found in other conditions, hemorrhages that are multiple, involve more than one layer of the retina, and extend to the periphery are very suspicious for abuse. The mechanism is likely repeated acceleration-deceleration due to shaking. Traumatic retinoschisis points strongly to abuse.

There are other causes of retinal hemorrhages, although the pattern is usually different than seen in child abuse. After birth, many newborns have them, but they disappear in 2-6 wk. Coagulopathies (particularly leukemia), retinal diseases, carbon monoxide poisoning, or glutaric aciduria may be responsible. Severe noninflicted direct crush injury to the head can rarely cause an extensive hemorrhagic retinopathy. Cardiopulmonary resuscitation rarely, if ever, causes retinal hemorrhage in infants and children; if present, there a few hemorrhages in the posterior pole. Hemoglobinopathies, diabetes mellitus, routine play, minor noninflicted head trauma, and vaccinations do not appear to cause retinal hemorrhage. Severe coughing or seizures rarely cause retinal hemorrhages that could be confused with AHT.